I didn't find an answer skimming: are they actively deleting old good subtitles and replacing with low-quality ones (as the title seems to suggest), have they simply changed their process for new subtitles, not spending as much to make nice ones? If they're actively deleting old good ones, that seems malicious.
I did find instances of them actively deleting old subtitles and replacing them with lower quality ones in the back catalog, yes. Which seems indicative of Crunchyroll wanting to eventually get rid of good subtitles altogether.
What possible incentive could they have for doing this?
I worked at crunchyroll.
Keeping the "hard subs" content is a lot of videos as the subtitles were encoded into the video stream.
This makes CDNs and other systems more difficult to utilize because we have a ton of video streams with just caption changes as opposed to just the Japanese audio source + caption files.
It's one of those things that doesn't seem that problematic till you include all the video_qualities to support streaming bandwith. So you also get a #hardSubLanguages * #videoQualities
Obviously you probably thought about it but what about rendering the subtitles on top of the video stream? Was there a reason it was not possible (e.g DRMs?)
This kind of softsubbing is what Crunchyroll primarily does, but it has hardsubbed encodes for devices that cannot do softsubbed rendering of the ASS subtitles that Crunchyroll uses. I go over some ways in how they could do away with these hardsubbed variants in the article without any notable loss in primary experience quality.
They could borrow a trick from Netflix mentioned elsewhere in this thread: https://netflixsubs.app/docs/netflix/features/imgsub
I’m pretty sure it’s not too hard to implement an ASS → PNG renderer (especially considering vibe coding is now a thing). Then, just need to split out subs that can be actual text somehow from the ones that have to be overlays.
Apart from that... surely they could at least keep ASS subs for the players that support it, and serve “fallback” subs for low-end devices?
ASS can have frame-by-frame animation IIRC, so a stream of PNGs could end up being quite high bitrate with high complexity files
It can, but that doesn't mean they use that functionality.
[dead]
It is harder than you think and will break on many more devices than you think.
So you make the business decision to stop supporting weird devices that can't do the job right? Why on earth does a cartoon streaming site need provably-correct subtitle support for devices that clearly suck?
Because the owners of those devices are paying them.
I’ve mentioned it elsewhere, but... why not keep proper ASS subs and fallback subs for those devices?
If you hardsub the video, then you need to have a full copy of the video for every language. That's the opposite of what people want. They want a single textless video source that can then accommodate any internationalization.
The article claims that you can slice up the video and only use language-specific hardsubs for parts that need it. I'd be interested if there are technical reasons that can't be done.
To be more specific, basically all online streaming today is based around the concept of segmented video (where the video is already split into regular X-second chunks). If you only hardsubbed the typesetting while keeping the dialogue softsubbed (which could then be offered in a simpler subtitle format where necessary), you would only need to have multiple copies of the segments that actually feature typesetting. Then you would just construct multiple playlists that use the correct segment variants, and you could make this work basically everywhere.
You can also use the same kind of segment-based playlist approach on Blu-ray if you wanted to, though theoretically you should be able to use the Blu-ray Picture-in-Picture feature to store the typesetting in a separate partially transparent video stream entirely that is then overlaid on top of the clean video during playback.
Technically it's possible.
We did do inlaid server-side ads that way for a while.
IT just takes an excessive amount of work.
The real solution is just the full support of ASS/TTML/VTT subtitles on all platforms. Usually smart devices are kind of only partially supported.
For instance - casting to a chromecast fallsback to SRT.
It's incredibly fragile at the CDN level if deployed at scale for a start.
You'd see playback issues go up by 1000%.
In the nicest possible way, it is pretty clear that this article was written by somebody who has only ever looked at video distribution as a hobbyist and not deploying it at scale to paying customers who quite reasonably get very upset at things not working reliably.
What would be the problems? When I’ve looked into streaming video before (for normal, non ripping reasons), I’ve noticed that most are already playlists of segments. You’d just need to store the segments that are different between versions, which should be better than keeping full separate versions which is what they apparently do currently.
This is just an excuse. There needs to be a hard english sub and then keep other languages can be single video with different text file. Deleting 80% good things only to keep other 20% happy should not be an excuse.
Only english is the most popular and just keep it. Most of the good hard subs are made for english and that is what people want.
That is exactly what I thought and I am not even a native English speaker. My English is infinitely better than my Japanese though, so if I cared about anime I’d much rather watch a good English version rather than a bad German one
Guess how well supported soft subs are on smartTVs etc? :)
It's really tough when you need to scale these things across 20 platforms.
What reason would there be for removing old, loved subtitles? Licensing fees?
Crunchyroll is currently using a subtitle rendering stack that is highly unique in the media industry, being based on the Advanced SubStation Alpha (ASS) subtitle format. It seems that the current executives would like to replace this unique stack with something more "industry standard" (and far less capable), but they can't do that as long as their back catalog is full of ASS subtitles - if they just switched stacks without doing anything else, all of these back catalog subtitles would just stop working completely. Which is why in order to perform such a stack switch, all old subtitles would need to be replaced with worse ones to make them compatible with the less capable new stack.
Per the article, on top of this, the Crunchyroll subtitle authoring tools and the Crunchyroll player both use ASS. So the directive to have the “master” copy be the more widespread and limited TTML… means that many new shows are doing ASS-to-TTML-to-ASS conversions! Quite literally the lowest common denominator of shared functionality.
Well the subtitles are ASS, alright, I'll give em that!
who in the world thought that a .ass file extension would be a good idea, sigh.
https://fileformats.fandom.com/wiki/SubStation_Alpha
From submitted article:
>Yes, it really is called ASS. Its predecessor, SubStation Alpha, went by SSA, and since these formats were developed by enthusiast hobbyists, I believe they couldn’t resist the siren song of a little bit of cheeky naming.
IIUC, it comes down to simplifying playback/subtitle rendering to the lowest common denominator among the various western streaming platforms.
The good/old subtitles in the ASS format required a more complex playback system than what Netflix/Hulu (and maybe blueray players) currently offer. This could be worked around by burning the subs into the video stream, but then you need to keep separate copies of your (large) video files for each subtitled language.
That doesn't seem like it'd be such a huge problem to me, but what do I know?
The post does a good job explaining the effective monopoly system at play that prevents real competition to provide any pressure to improve or maintain the prior quality.
It's an X*N*N problem: n_videos, bitrates, formats.
Assuming each video in its largest bitrate is... 2gb for example, and assuming S3 is $0.025/gb, that's a nickle per month or let's say $0.50/yr for that video.
Next up is reduced bitrates, assume you go from 2gb to 1gb and finally 500mb. Round up and you're at $1/video.
Now duplicate it to AV1 and MP4, and multiply that by English, French, and Spanish (oh, and let's say Japanese and Chinese too for good measure).
So a single 2gb video goes from $1/yr to $10/yr, and you're not doing "the dumb simple thing" for subtitles which would basically 4x your cost over "commodity subtitling services".
Or "simplify, simplify simplify", you reduce costs (cha-ching!), and become compatible for syndication or redistribution (cha-ching!)
... and they would have gotten away with it too if it weren't for those meddling kids!
S3 is >100x more expensive than hosting it yourself. You shouldn't argue about things being expensive using cloud prices.
Except ASS streams really aren't that big and don't have to be stored with each encode. They can just be in a separate file. And this is how cr used to serve them. Before they used hardware drm you could just download all of the separate sub tracks.
You don't need to multiply anything here, except the number of sub streams. One is ass, the other the primitive standards Netflix and other surges use.
Someone else was saying they maintained burn-in subs for devices that didn't handle ASS renderers. Even without accounting for the burn-in versions, using non-standard subs still bumps them off of commodity subtitling services and limits distribution/syndication.
Edit: and to the peer comment regarding S3 vs self host: regardless of 10x cloud cost, it's still 10x volume. Where 1TB local would do, now you need 10TB (10x the cost).
The labor of ass to ttml is there yeah. But the the factors are n_videos * languages * 2 Formats. And considering these are pretty compressible text(34MB->4MB for a completely bonkers sub track that includes animations, animated fonts and otherwise transformed text). I can't imagine that hosting costs cost more than their analysis.
Lots of old CR content was fan subs which may or may not have followed any kind of "best practices". It was also a reason that CR struggled with legitimacy with the studios in the early days. They tried to remove content once a US company licensed a show as way to show they weren't pirates, but the friction was always there.
Without knowing for sure, a conceivable strategy would be making it so that users who don't know that "good subs" can exist more readily accept "bad subs" as the standard. Same playbook as Google lobotomizing search and the index that search can access because that traditional search product is more efficacious than Gemini, which Google is trying to push on users.